Skip to main content

Systematic Reviews: Additional resources

Discipline span

When embarking on your initial research, consider whether it covers several disciplines. If it does, ask yourself the following questions: 

 

  • which disciplines?
  • what are my search criteria?
  • what could be other researchers' search criteria be?
  • which database/s and other resources should I use?
  • do I need to streamline my search?

 

If your research doesn't cover multiple disciplines, consider:

 

  • how specialised is my research?
  • which database/s and other resources should I use?
  • will I need to compromise/re-define my search terms?

 

It will be easier to get started if you think about how your research connects to a wider audience/research group. Other professions may approach the same area/s you are interested in from a different viewpoint and may use different search terms and phrases to the ones you may initially think of.

Remembering The Rules – Revisiting Boolean

This short online resource will highlight some of the common pitfalls with using Boolean logic to help you to avoid them in your own searches.

Download a printable version of this resource (PDF)

My Learning Essentials

To help you to get the most out of this guide, you may find it useful to explore the following online resources.

  

This set of online resources will demonstrate the process of performing a search on a number of database platforms.

View all workshops and online resources in this area on the My Learning Essentials webpages.

Knowing your sources

This online resource will explore four different health science databases and will demonstrate how to maximise your searches.

Knowing your resources

Glossary

Systematic review

A high-level overview of primary research on a particular research question that tries to identify, select, synthesize and appraise all high quality research evidence relevant to that question in order to answer it.

Critical appraisal

The assessment of evidence by systematically reviewing its relevance, validity and results to specific situations.

Synthesise

To combine so as to form a more complex, product.

Sensitivity vs precision

Literature searches that focus on a high level of sensitivity may return lots of results, but they will not necessarily all be relevant and appropriate for your research area. If your research is in an area that is fairly new, a high sensitivity search may bring up articles that are useful and relevant, because the net has been cast wider to catch more results. However, high sensitivity searches should be carried out with caution because the results could fall either way – you may get lots of relevant or irrelevant articles – or a mixture of the two. Searches that focus on more specific or precise terminology usually yield fewer results. A high proportion of these results may be relevant to your research but again use this method with caution: if you are too precise and narrow in your search criteria you could miss records that could have been useful – especially in topics that are relatively new research areas. In both cases, careful usage of search terms, search tools and limits, and accurate usage of truncation will determine how successful your searches will be. More detailed information about sensitivity versus precision can be found in the Cochrane handbook.

Citation pearls

A citation pearl is a term for a particularly significant piece of research. It may be especially valuable because it documents current (or previously held) “gold standard” viewpoints or practices. It will cite earlier sources of information and then be subsequently cited by other later research; all of these sources could be useful to your own research. Citation pearls can also get you thinking of your topic from other angles. The research area may be similar to yours but the search terms may differ and these could point you to other research avenues you had not previously considered. Databases such as Web of Science and the Ovid platform databases allow you to see how many times an article has been cited. You can also set up Cited Article Alerts in Web of Science. This is useful to keep up to date with newer sources, as well as tracing historical ones. However, you should be careful not to rely too much on the number of times an article has been cited; articles are cited for negative reasons as well as positive ones!

Meta analysis

A systematic method that takes data from a number of independent studies and integrates them using statistical analysis.

Qualitative

Relate to, measure, or measured by the quality of something rather than its quantity.

Quantitative

Relate to, measure, or measured by the quantity of something rather than its quality.

Grey literature

Research that is not published by a commercial publisher. This means it can sometimes be difficult to trace the research’s origin or the complete study. Grey literature can include conference proceedings, non-indexed journals (i.e. they are not on databases such as Medline), reports, and student dissertations and theses. Grey literature includes material such as theses, dissertations, conference papers and reports; these areas can provide other avenues for your research.Databases such as ProQuest Dissertations & Theses A&I will give details of theses and dissertations and other unpublished material. Some databases such as Scirus cover scientific and academic websites. Conference papers and reports by organisations such as the Department of Health can also direct you to resources you may not have considered. Manchester eScholar is a repository of all research outputs from the University of Manchester. The Library’s Research Services Team provides a dedicated service to support and advise users of Manchester eScholar.

Randomised controlled trial

A study usually involving at least two sample groups made up of people who share some basic criteria (eg same number of people in a group, same ratio of men and women etc.) who are randomly assigned to the sample groups. One group are given the treatment that is being trialled; the other group/s are given alternative interventions, which could be the standard treatment, a placebo or no intervention at all. The results of the trialled treatment are then compared against the results of the other groups at designated times and statistically assessed. RCTs are designed to reduce bias.

Quasi RCTs

Similar to randomised controlled trials but only partially random. This is usually due to some constraints, such as a more restricted age group, first hundred respondents to trial invitations etc. Quasi RCTs are useful in trying to establish a causal link between treatment results and patient groups when a pure RCT is not possible. However, a consequence of this is that the element of bias will be inevitably higher.

Empirical study

A study deriving from evidence presented based on experience and observation rather than that learned through theory.

Efficacy

“The ability to produce a desired or intended result” – Oxford Dictionary.

Meta-Synthesis

Meta-synthesis ”is the bringing together and breaking down of findings, examining them, discovering essential features, and, in some way, combining phenomena into a transformed whole”.

(Schreiber, R et al. Qualitative metasynthesis; Issues and techniques. In Morse, JM. Ed. 1997. Completing a qualitative project; details and dialogue. p314 Thousand Oaks, CA: Sage)

FAQs

Why do I need to search multiple databases?

It depends on your research topic. Some areas may be rich in resources in just one or two databases. However, for most detailed research projects such as systematic reviews you may need to scout around a bit, especially if your research is focusing on a new area of development.

Each database has its own strengths and weaknesses so you will need to experiment to find the best one for your topic. Some of the smaller database platforms have more limited functionality compared with the larger platforms, however the topic coverage may be very strong. In this case you need to be clear what you are looking for before using the database.

Most databases focus on a particular subject area such as medicine or law; you need to be aware of this if you are looking at an area spans across several disciplines. For example, if you are looking at the impact of long-term sectioning of a patient with mental health issues on the patient and their families, your search concepts will include mental health, social issues, human rights and law, so you will need to use health, law and sociology databases to consider all of these areas. The key databases for each discipline are listed on the Library’s subject pages.

There will always be a degree of duplication of results across databases. Abstracts are not unique to specific databases just as groceries are not unique to particular supermarkets. There will be some unique articles in a database just as there are some unique (own brand?) groceries in supermarkets.

Why should I not search multiple databases at once?

Different databases have different target audiences in mind, and the information within these databases is indexed according to its audience’s needs. For example PsycInfo is a psychology-focused database, so you would find more sub-categories relating to “attitude” than you would on Medline which is a more general index of mental health issues.

Cross-searching databases is a useful method of getting an overview of what literature is available, but you cannot drill down to the finer details.

For example, consider a search related to the attitude of health personnel towards patients who were terminally ill. If you searched Medline and PsycInfo together, the results from the two databases would merged and it would be difficult to separate them. This could be important because Medline results may be more focussed on keeping the terminally patient pain-free, whereas results from PsycInfo may be more focused on palliative and pastoral care.

For a systematic review, you will also need to be able to demonstrate which parts of your research fared well (or not so well) within specific results. Searching across multiple databases at once makes it much harder to do this; if you get a lot of results, it is difficult to work out which terms yield the most results.

What are the differences between the databases?

Many of the platforms have very similar features, but different platforms will present information in different ways.

Ovid and EBSCO are the most similar of the health sciences platforms. They also have the most comprehensive tools to help you find alternative words or phrases for your search concepts (these being MeSH and the Suggest Subject Terms respectively). It can be helpful to start with these platforms to help you to identify alternative search terms; you can them use these extra terms in other databases that do not offer the same facility.

It is important to look at smaller databases that don’t have the same level of functionality; with a little thought you can usually replicate your searches on these databases.

If you retain manageable segments within your search histories, these should be more easily replicated in other resources.

How do I explain my search? Do I say I searched Ovid/EBSCO/ProQuest, or name the specific database?

Ovid, EBSCO and ProQuest are platforms that each host a number of databases; it is more accurate to refer to the specific database. Some databases such as MEDLINE are available through several platforms, so it is also helpful to include details of which platform you’ve used to make it easy to replicate your search.

How can I be certain I’ve found everything?

You can’t. Things will always slip through the net and no database is 100% perfect. However, it is in your interests to try and obtain as much information as you can to ensure that your research is as comprehensive as possible. This may include searching relevant websites and grey literature in addition to standard database resources. When using sources like these, you need to be particularly careful to evaluate the information to ensure it’s of a high quality. Click here to learn more about evaluating your information.

Can I save all my results from different databases in one place?

Yes, using reference management software such as EndNote.

The Library runs EndNote workshops throughout the year, and we also have online support. Check out the referencing guide for more information.

Boolean issues

The Boolean operators AND, OR and NOT are common across search tools including databases and online search engines; they are used to group search concepts together.

Incorrect use of these operators are the most common cause of errors in searches, from basic formative searches to advanced systematic review searching. Be very careful with your search strings and ensure there are no discrepancies in meanings/definitions in your batches of terms. Check out the Revisting Boolen resource for more information.

Wildcard and truncation symbol usage

Truncation allows you to search for different variations on the same word but replacing the final characters of a word with a symbol. For example, searching for universit* would find both university and universities.

Wildcards allow you to search for multiple variations of a word at once by substituting one or more characters with a symbol. For example, searching for wom#n would find both woman and women.

The asterisk (*) has become the most common truncation symbol used at the end of a root term for most databases For example - medic* = medic, medication/s medical/s, medically, medicine/s etc..

However, there are differences across platforms in which symbols are used for wildcards and truncation, so it is important that you use the correct ones for the platform you are using.

It is important to check which database platforms use which symbols. Guidance on Truncation and Wildcard can usually be found in the database platform’s help pages.

If there are any variants in these symbols present in your search tables or search histories, it is advisable that you indicate this to ensure that there is no doubt over their correct usage.

Referencing

If you need help with referencing visit the Referencing Guide.

referencing guide

The guide gives an overview of referencing, including different referencing styles and referencing software.

Creative Commons Licence This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International Licence.

You can use this box to make notes on this page. Please contact us if you encounter any problems with this tool or if you have any feedback. Ensure you email your notes to yourself before you close your browser.

 

Save:
Save as text file:
Email notes to yourself:
>
Notes